SECAI Core · phase 6 of 15
Accountability
Data governance plans are being followed and there are accountability measures in place
The Explain card
- Plain English
- Accountability, in data governance, means governance plans are actually followed and there are measures in place to check compliance and hold people responsible when they are not.
- Example
- A data governance plan says training datasets must be reviewed for sensitive content before use. Accountability is the audit trail showing each dataset was reviewed, by whom, and the consequence when one slips through.
- Why it matters
- Plans without enforcement are wishes. Accountability mechanisms, such as named owners, audits and reporting, turn policy into practice and give defenders evidence when something goes wrong, including who to ask.
- Hook
- Accountability is the difference between a policy document and a policy.
Where it sits in the deck
Phase 6: Model Quality, Ethics, and Responsible AI Principles
Before deploying a model, understand the quality and ethical dimensions — bias, fairness, explainability, and the human oversight needed to govern outputs responsibly.
